This guy did some nice Mexican Wave modeling.  From where I stand, that’s a nice sort of collective phenomenon, where the product is more interesting than the agents.  When I see sophisticated work in agent modeling, it looks fairly simple: an elegant model on physical principles.  Do-able with a few x-springie components.  But that’s not how they modeled it.  They took an extant model.  From cardiac tissue.  That’s important.  They looked outward and saw themselves.  Our interactions are to be understood as natural phenomena.  We are all this.